status

Status: Unified System View

You are generating a system status snapshot — a single-command overview of backlog, recent activity, team health, and last session state. This is a read-only, low-cost diagnostic.

Focus area (optional): $ARGUMENTS

When to Use

• At the start of a session to orient quickly
• After returning from a break to see what changed
• When you want a full picture without dispatching agents or running multiple commands
• Before deciding what to work on next

Don't Use When

• You need deep diagnosis of why a session feels wrong — use /session-health instead
• You already know what to work on and just need to start — /status adds no value if the path is already clear
• You only need backlog info — check your task tracker directly rather than the full status view

How It Works

Read known paths → Check task tracker + git commands → Format sections → Suggest actions

No agents dispatched. No files written. Pure read + format.

---

Phase 1: Collect Data

Gather all data before producing any output. Read from the known paths defined in rules/memory-layout.md.

1a. Last Session

Read .claude/tackline/memory/sessions/last.md. If it does not exist, note "No previous session data."

1b. Backlog

Check your project's task tracker for backlog statistics, ready tasks, and epic progress. If no task tracker is configured, note "No task tracker configured."

1c. Recent Activity

Run:

git branch --show-current
git log --oneline -5
git status --short

1d. Team

If .claude/team.yaml exists, read it to get the team name and member list. Then for each member, check .claude/tackline/memory/agents/<name>/learnings.md:

• Line count (via Grep or Read)
• Whether the file exists at all

If .claude/team.yaml does not exist, note "No team configured."

1e. Recent Activity Summary

git log --oneline --since="7 days ago" | head -20

If no recent activity, skip this section.

1f. Live Agents (conditional)

Call TaskList to retrieve active task state. If TaskList returns results, collect:

• Running tasks: tasks with status running — capture agent name, task description
• Completed tasks: tasks with status completed that have not yet been reviewed (no corresponding commit or task close since completion)
• Idle agents: agents listed in .claude/team.yaml (if present) that have no running or recently completed task

If TaskList returns no results, skip this section entirely.

---

Phase 2: Format Output

Present all collected data in this structure. Do not add commentary or analysis beyond what is specified — keep it mechanical and cheap.

## System Status

**Branch**: [current branch]
**Date**: [today]

### Last Session
[Contents of .claude/tackline/memory/sessions/last.md — include verbatim, trimmed to first 20 lines if longer]
[Or: "No previous session data."]

### Backlog
[Task tracker stats, formatted]
[Ready tasks — show task titles, max 5]
[Or: "No task tracker configured."]

### Epic Progress
[Epic status — show each epic with completion %]
[Omit section if no epics exist or no task tracker configured]

### Recent Activity
[git log --oneline -5 output]

**Working tree**: [clean | N files modified/untracked]
[If dirty, show git status --short output]

### Team
[Team name and member table:]
| Member | Role | Learnings | Health |
|--------|------|-----------|--------|
| [name] | [role from team.yaml] | [N lines] | [see below] |

Health values (mechanical):
- **active**: learnings file exists, >0 lines
- **cold**: learnings file missing or empty
- **bloated**: learnings file >120 lines

[Or: "No team configured."]

### Live Agents
[Only present when TaskList returned results.]

| Agent | Task | Status |
|-------|------|--------|
| [agent name] | [task description] | running / completed (awaiting review) |

Idle agents (in team.yaml but no active or recent task): [list or "none"]

[Omit this section entirely when TaskList returns no results.]

### Suggested Actions
[Numbered list, derived mechanically from the data above]

---

Phase 3: Derive Suggested Actions

Generate 1-5 suggestions based purely on observed state. Use these rules:

Condition	Suggestion
Working tree is dirty	"Commit or stash uncommitted changes"
Ready tasks exist	"Work on: [top 3 ready task titles]"
Blocked tasks exist	"Unblock: [blocked task titles]"
Any team member is bloated (>120 lines)	"Run /retro to consolidate [member] learnings"
Any team member is cold	"Dispatch work to [member] to build learnings"
No last session file	"First session — review backlog and pick a starting task"
In-progress tasks exist	"Resume in-progress work: [task titles]"
Epics with all children complete	"Close completed epics in your task tracker"
Everything is clean and no ready tasks	"Backlog is clear — create new tasks or run /blossom to explore"
Completed tasks awaiting review (from TaskList)	"Review results from [agent name]: [task description]"
Idle agents with ready backlog tasks	"Dispatch [agent name] to: [top ready task title]"

Only include suggestions that match the current state. Do not invent conditions.

---

After the Suggested Actions section, emit a pipe-format summary so /advise can consume status signals:

## System status signals

**Source**: /status
**Input**: [focus area from $ARGUMENTS, or "full system status"]
**Pipeline**: (none — working from direct input)

### Items (4)

1. **Backlog state** — [N open tasks, N ready, N blocked, N in-progress]
   - ready: [top 3 task titles, or "none"]
   - blocked: [N tasks blocked, or "none"]
   - epics: [N epics, M% complete on average, or "none"]

2. **Recent activity** — [active | quiet | stalled]
   - branch: [current branch]
   - commits-last-7d: [N]
   - working-tree: [clean | N files modified/untracked]

3. **Team health** — [healthy | cold members | bloated learnings | no team]
   - members: [N total, N active, N cold, N bloated]
   - action-needed: [member names needing attention, or "none"]

4. **Suggested actions** — [N actions derived from state]
   - actions: [top 3 suggestions as short phrases]
   - confidence: CONFIRMED

### Summary

[One paragraph: current backlog shape, recent activity level, team health, and the single most important action to take next.]

Guidelines

1. Fast. This should complete in under 10 seconds. No agent dispatch, no synthesis essays.
2. Honest. Show actual data. "No team configured" and "No task tracker configured" are valid outputs.
3. Mechanical. Suggested actions are derived from rules, not LLM creativity. If no rule matches, say "No immediate actions suggested."
4. Read-only. No file writes, no bead creation, no git commits. Pure observation.
5. Graceful degradation. Every section works independently. Missing data sources produce a one-line fallback, not an error.

See also: /evolution (definition change tracking). /standup (team-aware equivalent — use standup when a team is configured). /session-health (focused diagnostic when something feels off in the current session).